How Often Did Segments Evolve?
Not only are the hypotheses on the evolutionary origin of segments subjects of heated debate, but also are the questions as to where and how often segments evolved. Depending on the perspective, single, twofold, or threefold independent evolutionary formation has been postulated (Balavoine and Adoutte 2003; Seaver 2003; Vellutini and Hejnol 2016). In most cases it is assumed that segmentation in chordates, annelids, and arthropods is homologous within each of the respective groups. The different hypotheses are based on varying answers to the question whether segments of these three major groups are homologous to each other and on their differing tolerance with respect to a repeated convergent loss of segmentation. Moreover, they reflect often unconscious or unclear concepts regarding segments.
For several years the hypothesis has been favored that segmentation evolved once and that the stem species of Bilateria (Urbilateria) was already segmented (see Balavoine and Adoutte 2003). Yet, recent analyses of bilaterian phylog- eny led to the view that the last common ancestor of annelids, arthropods, and chordates was not Urbilateria but the stem species of the large bilaterian subtaxon Nephrozoa (Cannon et al. 2016) (Figure 1.10). The putative sister group of Nephrozoa, the Xenacoelomorpha do not show any signs of segmentation. Hence, the discussion has been shifted to the question of whether segmentation evolved in the stem lineage of the Nephrozoa (Treffkorn et al. 2018). Due to similarities in the expression of segmentation genes, some authors view the segments of annelids, arthropods, and chordates as homologous (Carroll et al. 2001). According to this view, the stem species of the Nephrozoa was already segmented. In this case, we must assume a multiple evolutionary reduction of segmentation in the other animals, that is, the close relatives of the annelids, the arthropods, and the chordates (Figure 1.10). The same also applies if we assume a twofold development of segmentation, once in the common line of the arthropods and annelids and, independent of that, in the line leading to the chordates. It would thus be most parsimonious to assume a threefold independent origin of segmentation (Figure 1.10). This view corresponds to the conclusions drawn from the minor structural correspondence between the serial units characterized as segments in the three groups.
Also, one must ask if the term “segment” is at all appropriate for the observed phenomena. Does the previously described concept of segmentation take the given conditions into account and does it even allow us to ask the proper questions about the origins of segmentation?
The contradictions between the theories of the evolutionary formation of segments already provide an indication that the concept of segmentation might itself be problematic. This assumption is even reinforced by the following problems.
FIGURE 1.10 Greatly simplified diagram of one of several hypotheses on the phylogenetic relationships of Nephrozoa. Only the lines of the segmented animal groups are named. An S stands for the emergence of segmentation, a crossed-out S for its loss. Once the segmentation has already emerged at the base of the nephrozoans (red S), this implies an eightfold loss of segmentation. If assuming a dual evolution of segmentation in chordates and the last common ancestor of annelids and arthropods (blue S), the segmentation still undergoes a sevenfold loss. Segmentation evolving independently in the three lines of annelids, arthropods, and chordates (green S) would be the most parsimonious solution, since it assumes only three evolutionary steps.